Re: [ccp4bb]: Twinning problems

[Anastassis Perrakis <perrakis@nki.nl>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

> Eleanor is right.

Most likely yes (ccp4bb default mode).
But the real problem is that Roe did not provide any of the evidence that let him
believe that this was a twin, he rather provided the evidence that showed it was NOT
a twin. Well, I would tend to assume that some other evidence (other than 'lack' of
anomalous signal) led him to believe that this was a twin ? Or is twinning
identification turning to be something like witch-hunting in the middle ages ?

> We solved a P42212 MAD structure using data from a
> crystal that only went to 3.5 A.  (We did have better native).  There
> was no evidence for an anomalous signal except possibly in the  low
> resolution bins (~15-8).

But if you solved it, there was anomalous signal. Thus, the criteria we use to pick
it up are inadequate.
Before throwing them to a ceremonial fire though, did you look at the normal
probability plots that SCALA produces ? Or to one of the simple-but-very-smart
criteria implemented to XPREP (well, not ccp4 prg, but ...) Btw, the XPREP criteria
(I(dano)/sigI(dano), correlation between ano1/ano2, anomalous/dispersive signals) are
not so hard to implement to e.g. SCALA or SCALEIT. Would that be a good thing to do
(or I should simply RTFM since they are already there ?) ?

    Tassos